A Nobel Approach to Detect Edge in Digital Image Using Fuzzy Logic

Edge detection is a popular technique to find out the boundaries of different objects in a digital image. These edges are searched on the basis of gradients that are available in the image. The gradients depend upon the intensity and value of pixels. In this paper, a new technique is proposed for edge detection which is based on the fuzzy rule-based system. Fuzzy-based system requires mainly three steps, conversion of inputs to fuzzy linguistic variables, then a set of rules that can be applied on inputs and final step is again converting back from fuzzy logic output to crisp output. Horizontal and vertical gradient vectors have been considered as inputs for the fuzzy system. Gaussian membership functions and triangular membership functions are used for converting crisp inputs into fuzzy input. For defuzzification purpose centroid method is applied. The Mamdani model has been used to develop the system. The result of proposed system is better than most of the popular conventional edge detection techniques. The proposed system also minimizes the noisy details and can also be used variety of images.